# The brain changes its immune guard after 50 In the hippocampus of 40 donated brains, resident microglia thin out from about age 50 to 75. Cells with a blood-lineage signature appear in the same window. DNA methylation suggests their origin. The evidence comes from tissue taken after death, and the study did not track memory. ## What the paper shows, and where it stops. - Seen in tissue: In post-mortem hippocampus, resident microglia thin out between about 50 and 75. A second population appears, with stronger inflammatory signals. - Origin suggested by methylation: The new population shares epigenetic traits with blood monocytes. The study did not trace these cells inside a living person. - Disease hypothesis: The authors propose that the shift may contribute to chronic inflammation in neurodegenerative disease. The paper did not measure dementia in these donors. - Still open: We do not yet know whether the shift appears in living people over time or affects memory. ## Two populations in one hippocampus. Microglia are the brain's resident immune cells. They arise in the embryo and, in the classical model, renew themselves locally throughout life. In these 40 hippocampi, resident microglia thin out after age 50. ### The old guard Resident microglia arise in the embryo. They clear debris, monitor synapses and respond to danger signals. The paper finds fewer of them between about 50 and 75. ### The new guard In their place, cells appear with a stronger inflammatory program and traits resembling blood monocytes. The team interprets the methylation pattern as a change in cell origin. ## Gene expression and methylation answer different questions. Nathan Zemke, the first author, explains the distinction. Gene expression shows what a cell is doing now, while epigenetic signatures preserve clues about where it came from. The upper band mixes the points. Methylation separates them into two families below. > Gene expression tells us what a cell is doing today, but epigenetic signatures preserve information about where a cell came from. > > Nathan R. Zemke, in the NIH recap of the paper ## 40 hippocampi, ages 20–95. Each mark is one donor; positions are not real ages. Table S1 lists each case's age, sex and inclusion criteria. The Science PDF was unavailable at the date of this edition, so this grid does not copy the table. ## The blood-brain barrier also appears to thin in the same years. The team also saw fewer cells that support the blood-brain barrier. The public releases give no effect size. The barrier appears less continuous in the same interval as the guard changes. ## The evidence stops at cell origin. ### Seen in tissue Researchers analyzed 40 hippocampi from donors aged 20 to 95, using tissue taken after death. ### Cell origin suggested by methylation Methylation offers clues about cell origin. The study did not trace these cells inside a living person. ### Inflammation hypothesis The new population carries a proinflammatory program. The study did not measure a cause of disease, so the link remains a hypothesis. ### Living people, over time No longitudinal study has followed the same shift in living people. ### A test There is no validated clinical biomarker for this change yet. ### A treatment No therapy, supplement or protocol has been tested for this change. ## Four things we still do not know. ### Why are resident microglia lost? The paper describes the loss. The public releases do not say whether cells die, become exhausted or leave the tissue. ### Do blood cells enter, or do local cells reprogram? Methylation points toward a blood-like origin. A separate Stanford study, published later in Nature, describes immune cells entering from the periphery. Its design differs, and it does not confirm the 40 hippocampi in the Zemke paper. ### Does the shift appear outside the hippocampus? The study measured the hippocampus. To generalize to the whole brain, researchers need to study other regions. ### Does it predict cognitive decline in the same people? The public releases describe the donors as having no known neurological disease. The study did not track their memory over time. ## The sources behind the claim. The Zemke paper and releases from the NIH and New York Genome Center support the cell findings. The diagrams are editorial and do not reproduce figures from the paper. - [Zemke et al. · Science, 23 July 2026](https://doi.org/10.1126/science.adt8307) — Epigenetic and 3D genome reprogramming during the aging of the human hippocampus. Original research article. DOI 10.1126/science.adt8307. The PDF was unavailable at the date of this edition. - [NIH · Brain immunity may undergo a major midlife overhaul](https://www.nih.gov/news-events/news-releases/brain-immunity-may-undergo-major-midlife-overhaul) — The NIH release gives the 50–75 window, the sample of 40 and the replacement reading of microglia. - [New York Genome Center · 23 July 2026](https://www.nygenome.org/news-events/news/new-study-revises-understanding-of-brain-immune-cells-during-human-aging/) — Bing Ren's institutional release confirms the decline of embryonic microglia, the blood-like signature and the barrier finding. - [ScienceDaily · NIH recap, 6 August 2026](https://www.sciencedaily.com/releases/2026/08/260806050700.htm) — This recap repeats the NIH release, including Zemke's quotation and the author list. It adds no independent measurement. - [Ginhoux et al. · Science, 2010](https://doi.org/10.1126/science.1194637) — The 2010 reference for the embryonic, yolk-sac origin of microglia. The 2026 paper revisits this model in older adults. - [Stanford / Knight Initiative · Nature, 30 July 2026](https://news.stanford.edu/stories/2026/08/aging-brain-immune-system-research) — A separate study of immune cells entering the aging brain from the body. It uses a different design and journal, and does not confirm the 40 hippocampi in the Zemke paper. ## After 50, the hippocampus may change its guard. We still do not know what this change means for memory. A public edition by Marius Comper · Updated 14 August 2026 · Evidence searched through 12 August 2026